Biochemistry 1987, 26, 52-59
52
Enzymatic Methylation of Band 3 Anion Transporter in Intact Human Erythrocytes Lillian L. Lou* and Steven Clarke* Department of Chemistry and Biochemistry and Molecular Biology Institute, University of California, Los Angeles, California 90024 Received May 19, 1986; Revised Manuscript Received August 27, 1986
Band 3, the anion transport protein of erythrocyte membranes, is a major methyl-accepting substrate of the intracellular erythrocyte protein carboxyl methyltransferase (S-adenosyl-L-methionine: protein-D-aspartate 0-methyltransferase; E C 2.1.1.77) [Freitag, C., & Clarke, S. (1981) J. Biol. Chem. 256, 6102-61081. The localization of methylation sites in intact cells by analysis of proteolytic fragments indicated that sites were present in the cytoplasmic N-terminal domain as well as the membranous C-terminal portion of the polypeptide. The amino acid residues that serve as carboxyl methylation sites of the erythrocyte anion transporter were also investigated. 3H-Methylated band 3 was purified from intact erythrocytes incubated with ~-[methyl-~H]methionine and from trypsinized and lysed erythrocytes incubated with Sadenosyl-L- [methyl-3H]methionine. After proteolytic digestion with carboxypeptidase Y ,D-aspartic acid p-[3H]methyl ester was isolated in low yields (9% and 1%, respectively) from each preparation. The bulk of the radioactivity was recovered as [3H]methanol, and the amino acid residue(s) originally associated with these methyl groups could not be determined. No L-aspartic acid P-[3H]me'thyl ester or glutamyl y[3H]methyl ester was detected. The formation of D-aspartic acid /3-[3H]methyl esters in this protein in intact cells resulted from protein carboxyl methyltransferase activity since it was inhibited by adenosine and homocysteine thiolactone, which increases the intracellular concentration of the potent product inhibitor S-adenosylhomocysteine, and cycloleucine, which prevents the formation of the substrate S-adenosyl-L[methyl-3H]methionine. ABSTRACT:
R o t e i n carboxyl methyltransferases are widely distributed in nature and catalyze the formation of protein methyl esters. For example, a specific bacterial methyltransferase is responsible for the methylation of several membrane chemoreceptors on L-glutamyl residues, and these reactions appear to regulate the sensory output of these proteins (Stock et al., 1985a,b). Protein carboxyl methyltransferasesare also present in many mammalian tissues. The physiological roles of these eucaryotic enzymes are poorly understood, although it has been hypothesized that these reactions can also regulate the function of the various methyl-accepting substrates (Gagnon & Heisler, 1979; O'Dea et al., 1981). We have studied eucaryotic protein carboxyl methylation in human erythrocytes [for a review, see Clarke (1985)l. Several major methylated proteins have been identified in the membrane fraction of intact cells including the cytoskeletal proteins band 2.1 and band 4.1, as well as band 3, the anion transport protein (Freitag & Clarke, 1981). A methylated residue has been isolated in low yield from proteolytic digests of unfractionated methylated erythrocyte membranes and has been identified as the unnatural D isomer of aspartic acid @-methylester (McFadden & Clarke, 1982; Clarke et al., 1984). Studies using synthetic peptides as methyl acceptors
'This research was supported by National Institutes of Health Grants GM-26020 and EY-04912and by a Grant-in-Aid from the American Heart Association, with funds contributed in part by the Greater Los Angeles Affiliate. L.L.L.was supported in part by U S . Public Health Service Training Grant HL-7386. *Address correspondence to this author at the Department of Molecular Biology, Princeton University, Princeton, NJ 08544. 'Present address: Department of Pharmacology, Stanford University, Stanford, CA 94305.
0006-2960/87/0426-0052$01.50/0
have indicated that a second methylated product is L-aspartic acid a-methyl ester (Murray & Clarke, 1984; Aswad, 1984). The origin of the substrates for these reactions (L-isoaspartyl and presumably D-aspartyl residues) in cells is uncertain, although they may arise as a consequence of spontaneous racemization or deamidation events. On the basis of these data, it apgears that the methylation reaction may be involved in the metabolism of age-modified proteins. In the present work, we have focused on delineating the sites of methylation of the band 3 anion transport protein in human erythrocytes. The structure and function of this protein have been reviewed recently (Macara & Cantley, 1983; Jay & Cantley, 1986). Band 3 is a major endogenous substrate for the human erythrocyte protein carboxyl methyltransferase in both intact (Freitag & Clarke, 1981) and broken cell preparations (Terwilliger & Clarke, 1981). On the basis of inhibitor studies, this methylation reaction does not appear to directly regulate the anion transport function of this protein (Lou & Clarke, 1986). Band 3 is an ideal candidate for these experiments because previous work has established the topography of this protein within the bilayer [see, for example, Jennings & Nicknish (1 984)] and functional domains have been identified for anion transport (Grinstein et al., 1978; Ramjeesingh et al., 1980), cytoskeletal association (Bennett & Stenbuck, 1979; Bennett, 1982), and binding of cytosolic proteins (Murthy et al., 198 1). Partial sequence information is available for the human band 3 protein (Mawby & Findlay, 1982; Brock et al., 1983; Kaul et al., 1983), and a complete sequence of the mouse protein has recently been proposed (Kopito & Lodish, 1985). In this paper, we report on the identification of a methylated residue and the localization of the major methylation sites in band 3 of membranes isolated from erythrocytes incubated 0 1987 American Chemical Society
BAND 3 METHYLATION I N ERYTHROCYTES
with ~-[methyl-~H]methionine. These probably represent the endogenous sites, since cells were incubated under conditions of physiological pH, temperature, and methionine concentration. EXPERIMENTAL PROCEDURES Materials. L- [methyL3H]Methionine and S-adenosyl-L[methyl-3H]methionine were purchased from Amersham. Adenosine, L-homocysteine thiolactone, cycloleucine, Lmethionine, S-adenosyl-L-homocysteine, phenylmethanesulfonyl fluoride, penicillin G, streptomycin sulfate, and all proteases were purchased from Sigma. L-Aspartic acid pmethyl ester hydrochloride was purchased from Vega Biochemicals. D-Aspartic acid @-methylester hydrochloride and L-leucine N-carboxyanhydride were prepared as described (McFadden & Clarke, 1982). Reagents for gel electrophoresis were obtained from Bio-Rad. Reagents and buffers for amino acid analysis were obtained from Pierce. All other reagents were analytical reagent grade. Preparation of Erythrocytes. Human erythrocytes were prepared from freshly drawn blood as described (Barber & Clarke, 1984). Carboxyl Methylation of Erythrocyte Membrane Proteins. Intact cells were incubated under the following conditions. An aliquot of ~-[methyl-~H]methionine (72-93 Ci/mmol, 13.9-10.8 pM, 3.5 volumes) was lyophilized to dryness in a polycarbonate centrifuge tube. It was resuspended in 2.5 volumes of isotonic phosphate buffer (118 mM sodium phosphate, pH 7.4) containing 16.7 mM glucose prior to the addition of 1 volume of washed erythrocytes. Incubation was carried out at 37 "C with shaking for 3 h (28.6% hematocrit, final concentration of 10.8-13.9 pM ~ - [ m e t h y l - ~ H ] methionine). For methylation-inhibited samples, either 5 mM each of adenosine and homocysteine thiolactone or 50 mM cycloleucine was included in the incubation buffer. At the end of the incubation period, cells were diluted with at least 30 volumes of ice-cold 118 mM sodium phosphate, pH 7.4, and then collected by centrifugation. Pelleted cells were hypotonically lysed by using at least 30 volumes of 5 mM sodium phosphate, pH 8 at 4 OC. Membranes were pelleted by centrifugation at 4 "C (26000g, 6 min). The supernatant and "button" underneath the soft membrane pellet were removed by aspiration. The membranes were washed in the same buffer at least twice more. Band 3 was 3H-methylated in lysed cells by the following procedure. Washed intact erythrocytes (450 pL) were fractured by freezing in a dry ice/2-propanol slurry (-70 "C) prior to addition of 37 pL of S-adenosyl-L-[methyl-3H]methionine (15 Ci/mmol, 66.7 pM). Lysed cells were allowed to incubate for 1 h at 37 OC and then resuspended in 30 volumes of ice-cold 5 mM sodium phosphate, pH 8. Membranes were washed as described above. Separation of Extrinsic and Intrinsic Erythrocyte Membrane Proteins. Extrinsic membrane proteins including bands 1, 2, 2.1, 4.1, 5, and 6 and hemoglobin were removed from the intrinsic membrane proteins (mainly band 3 and the PAS bands) by extraction with 10% (v/v) acetic acid as described (Steck et al., 1976). The membrane pellet was resuspended in 10 volumes of 10% acetic acid and centrifuged at 27000g for 20-25 min at 4 OC. It was found that the extraction became more efficient when the suspension was allowed to incubate at room temperature for 10 min prior to centrifugation, and an additional wash with 100 volumes of 10% acetic acid was done. The resulting pellet, which contained the intrinsic proteins, was typically more tightly packed, reduced in volume, and transparent.
VOL. 26, NO.
1, 1987
53
Proteolytic Fragmentation of Band 3: (A) Intact Cells. Trypsin and chymotrypsin digestions of erythrocyte external membrane polypeptides were performed according to procedures described by Drickamer (1976). Packed erythrocytes were incubated at 50% hematocrit in isotonic phosphate buffer with trypsin (from bovine pancreas, diphenylcarbamyl chloride treated, Sigma type XI, final concentration 19 pg/mL) for 3 h at 27 OC. To quench the protease digestion, phenylmethanesulfonyl fluoride in 95% ethanol was added to a concentration of 60 pg/mL. The sample was allowed to react for 30 min at 4 "C before cells were washed 3 or 4 times with 30-60 volumes of isotonic phosphate at 0 OC. Chymotrypsin digestions were carried out in phosphate-buffered saline at a 50% hematocrit. Washed erythrocytes were added to an equal volume of 2 mg/mL a-chymotrypsin (from bovine pancreas, tosyl-L-lysine chloromethyl ketone treated, Sigma type VII). The digestion was allowed to continue for 1-2 h at 24 OC. Phenylmethanesulfonyl fluoride was added to give a final concentration of 375 pg/mL, and the cells were processed as described above for the trypsin digestion. In all experiments, control cells were processed in parallel with the proteasetreated cells, except that the proteases were omitted during the incubation. Carboxyl methylation of membrane proteins was carried out immediately following fragmentation. (B) Membranes. Erythrocyte membranes were digested with trypsin after the procedure of Steck et al. (1978). Trypsin (40 pg/mL) in 5 mM sodium phosphate, pH 7.2, was added to an equal volume of membranes (4-6 mg of protein/mL). The digestion was allowed to continue for 1 h at 24 OC. The reaction was quenched by the addition of 10 volumes of 10% acetic acid, and membranes were extracted as described. Chymotrypsin digestion was performed as described by Ramjessingh et al. (1980) and Rao and Reithmeier (1979). Membranes were added to an equal volume of 5 mM sodium phosphate, pH 7.2, containing 0.1-4 mg of chymotrypsin/mL. The incubation was performed at 37 OC for 1 h, and the reactions were quenched as described above for the trypsin digestion. Separation of Polypeptides by Sodium Dodecyl Sulfate/ Urea-Polyacrylamide Gel Electrophoresis. Slab gel electrophoresis was performed in the pH 2.4 buffer system described by Fairbanks and Avruch (1972) and O'Connor and Clarke (1985) except that samples were denatured in buffer containing 4% sodium dodecyl sulfate (SDS)' and 4 M urea and the gels contained 6 M urea. Quantitation of Methylation of Proteolytic Fragments Separated by Polyacrylamide Gel Electrophoresis. Protein methylation is measured by the radioactivity incorporated into specific polypeptide bands on SDS gels. Densitometric traces of the Coomassie-stained bands in wet gels and the film resulting from fluorography (OConnor & Clarke, 1985) of the same processed gel were made by using a Helena Quick-Scan R & D instrument. The number of [3H]methyl esters in polypeptide bands on gels was quantitated from the area of the densitometric trace of the film, and the number of polypeptides was calculated in a similar way from the densitometric trace of the Coomassie-stained gel and knowledge of the molecular weight of the protein species. Base lines for individual polypeptide bands were drawn in the same way for both the film and Coomassie profiles. Isolation and Determination of Stereoconfiguration of Aspartic Acid p-[3H]Methyl Esters from Band 3 Enriched Abbreviations: SDS, sodium dodecyl sulfate; AdoHcy, Sadenosylhomocysteine; AdoMet, S-adenosylmethionine; L-LeuCA, Lleucine N-carboxyanhydride.
54 B I O C H E M I S T R Y
LOU AND CLARKE
N
-
ZZK-Z~KII)
IBK-ZOKII)
0
52K-60K lml
b
T,
T2
55K-59KIrn
I
C
9 0 K - l 0 0 K rn
40K-41K 1s)
35K-38KM
..
C
,d
Ch, 14.5K-19Kh)
Chz 13.5K-15Klrnl
Ch,
7K-9Klrnl
Ch,
e
Chg
FIGURE 2 Major proteolytic cleavage sites of the band 3 polypeptide. N, N-terminal; C, C-terminal, K,XlO’ daltons; Ch, chymotrypsin cleavage site; T, trypsin cleavage site. The suffixes s and m indicate
soluble and membrane-bound fragments, respectively. Each cleavage site, as indicated by a breakage in the polypeptide, represents a small region of amino acid sequence that is sensitive to that particular protease. Stippled fragments represent methylated species. (a) No protease, or trypsin digestion of intact cells;(b) mild trypsin digestion of membranes; (c) chymotrypsin digestion of intact cells; (d) mild chymotrypsin digestion of membranes; (e) extensive chymotrypsin digestion of membranes.
,L,, ........ i,
,
coom(J58ie-s1oi” , , ._._.... 1,. ..,..........................................
Rodiomlivily
1
2
4
6
8
II
M i g r a t i o n (cm)
FIGURE 1: Separation of peripheral and integral membrane proteins by acetic acid extraction. Membranes were isolated from erythrocytes incubated with ~-[methyl-’H]methionineand extracted with 10%
acetic acid as described under Experimental Proccdum. The extracted supernatant and the residual pellet samples were analyzed by electrophoresis on a 7.5% polyacrylamide gel wntaining 1% SDS and 6 M urea. A densitometric trace of Coomassie blue stain (dashed line) is shown superimposed on a densitometric trace of a fluorograph of the same lane (solid line). The top panel represents 40 @gof total membrane protein before extraction (membranes from about 5.7 X IO’ cells). The middle panel represents the supernatant,and the bottom panel represents the residual pellet after extraction of a similar amount of membranes. The migration positions of molecular weight standards are indicated by the arrows in the top panel. The band of residual hemoglobin is labeled Hb.
Membranes. Membranes (2-6 mg of protein/mL) prepared from trypsinized erythrocytes incubated with S-adenosyl-L[methyl-’Hlmethionine under lysis conditions, and acetic acid extracted membranes from intact erythrocytes incubated with ~-[methyl-’H]methionine, were digested with equal volumes of carboxypeptidase Y (bakers’ yeast, 2 mg of protein/mL, in pH 5 citrate buffer, Sigma) in the presence of 1% (v/v) Triton X-100 at 37 OC for 15-21 h. The digestion products were analyzed by cation-exchange and gel filtration chromatography as described in detail previously (Clarke et al., 1984). Stereoconfiguration of the radioactive material that coeluted with an added standard of aspartic acid &methyl ester was determined by the diastereomeric dipeptide method of Manning and Moore (1968). The resulting L-leucyl dipeptides were separated by ion-exchange chromatography as described (McFadden & Clarke, 1982: Clarke et al., 1984). Quantitation of Methyl Ester Formation in Band 3. Protein-associated methyl esters were defined as acid-precipitable, base-labile, alcohol/tolueneextractahle, and volatile radioactivity. This is measured by the assay procedure described previously (OConnor & Clarke, 1983). As assay performed with whole membranes provides a value of the picomoles of methyl ester in the membrane per milligram of total membrane protein. This value is multiplied by the percent of band 3 associated radioactivity relative to total protein-associated radioactivity to give the picomoles of methyl ester in band 3 per milligram of total membrane protein. The percent of radioactivity in whole membranes contributed by band 3 is defined as an area under the radioactive peak, in a densito-
metric trace of a fluorograph, comigrating with the band 3 Coomassie stain in the same lane divided by the area under all the radioactive peaks in the same densitometric trace. RESULTS Preparation of Isotopically Pure Band 3 by Acetic Acid Extraction of Intact Cell 3H-Methylated Membranes. When intact erythrocytes are incubated with ~-[methy/-’H]methionine, the major methylated species are bands 2.1,3, and 4.1. The data in Figure 1 show that it is possible to extract the non-band 3 associated radioactivity from the membrane with acetic acid. The membrane pellet prepared in this fashion can then be used as a source of ’H-methylated band 3. Sites of Carboxyl Methylation on Band 3 Separation of the Cytoplasmic Domain and the Transmembrane Domain by Proteolytic Digestion. Band 3 is composed of two functionally and structurally dissimilar domains: a hydrophilic cytoplasmic portion (N-terminal) that associates with cytoskeletal and cytosolic proteins and a hydrophobic transmembrane portion (C-terminal) that transports anions across the bilayer. Mild trypsin digestion of unsealed membranes results in the release of the cytoplasmic domain of hand 3, leaving the transmembrane domain in the membrane. The N-terminal-containing cytoplasmic domain consists of a water-soluble 41 000-dalton fragment, which is readily cleaved to yield two major polypeptides with molecular weights of 22CQ0-23000 and 16OOC-20000 (Figure 2: Steck et al., 1978; Reithmeier, 1979). The membrane-bound fragment has an apparent molecular weight of 52000, and it characteristically migrates as a diffuse band in a polyacrylamide gel (Ramjeesingh et al., 1980; Jennings et al., 1984). When ’H-methylated membranes were digested with trypsin and extracted with acetic acid as described above, the radioactivity and Coomassie stain of band 3 disappeared, and a new peak of radioactivity comigrated with a 60000-dalton transmembrane fragment (Figure 3). This fragment was found to contain about 60% of the radioactive methyl groups associated with the intact band 3 polypeptide in a control sample incubated without protease. The rest of the methyl groups would be expected to he in the soluble fraction of the digest including the two major soluble fragments (16 000-20 000 and 22000-23 000 daltons). The smaller fragment, which is sensitive to further digestion, is recovered invariably in low yield while the larger fragment can be obtained in good yield (Steck et al., 1976, 1978). Analysis of the supernatant after
VOL. 26, NO. 1, 1987
B A N D 3 M E T H Y L A T I O N IN E R Y T H R O C Y T E S
:\ ::
94K 68K
4 4
bo!;
43.5K 29.5K
4
+
21K
4
-
14K
94K68K
1
4
I
43.5K295K
1
21K
4
55
14K
4 4
Confro/
3
Control
1 :
h
c .In
C
e, +
-C
u .-
c I 0) c E, ._ u) C
e,
0
Chymotrypsin-treated
0
2
4
6
8
10
Migration (cm) FIGURE 3: Removal of cytoplasmic domain of band 3 in erythrocyte membranes by trypsin digestion. Membranes isolated from 3Hmethylated intact cells were digested as described under Experimental Procedures. A control sample was incubated in parallel except that the protease was omitted. Membranes were extracted with acetic acid, and the residual pellets were analyzed by electrophoresis. The protein content and the corresponding radioactivity were detected and presented as described in the legend of Figure 1. In this particular experiment, the control lane represents a sample of 60 pg of total membrane protein (equivalent to 8.6 X lo7cells), and the trypsin lane represents a sample of digested membranes from an equivalent of 21 X lo7 cells (2.4-2.5 times the amount in the control sample). The radioactivity recovered in the transmembrane trypsin fragment is 1.4 times the amount in the control sample. Therefore, after correction for the amount in the control sample materials applied according to the cell equivalence, the recovery of radioactivity in the proteolytic product was found to be 57% of that in the intact band 3 polypeptide (see Experimental Procedures for details). From the results of four separate experiments, the radioactivity of the 60 000-dalton transmembrane fragment was calculated to be 59% 3% of that of intact band 3 in an undigested sample. The value is expressed as mean standard deviation.
*
*
digestion revealed radioactivity in the 22 000-dalton N-terminal fragment (data not shown). Proteolytic Fragmentation of Intact Erythrocytes Followed by Incubation with L- [methyl-3H]Methionine. When intact erythrocytes are exposed to chymotrypsin, band 3 is cleaved into two major membrane-bound fragments with molecular weights of approximately 59000 and 38 000 (Steck et al., 1978; Figure 2). The larger fragment contains the N-terminus and is easily detected in a polyacrylamide gel as a sharp Coomassie-staining band. The smaller fragment represents the C-terminal portion which migrates as a broad band in a polyacrylamide gel and stains poorly with Coomassie brilliant blue (Drickamer, 1976; Markowitz & Marchesi, 1981). Figure 4 shows that 3H-methylated polypeptides migrating at 59 000 and 39 000 daltons were generated by chymotrypsin digestion of intact cells followed by incubation with L[methyl-3H]methionine. Similar results were obtained when erythrocytes were incubated with L- [methyl-3H]methionine prior to chymotrypsin treatment. Approximately 60% of the radioactivity was found in the N-terminal 59 000-dalton fragment in each case. Taken together with the quantitation from the trypsin experiments, these results suggest that methylation sites occur in at least three regions of the band 3 polypeptide; about 40% of the methyl groups are located in the N-terminal41000-dalton fragment, 40% in the C-terminal 39 000-dalton fragment, and 20% in the central approximately 20 000-dalton region. Localization of Methylation Site(s) in the Chymotrypsin 59 000-Dalton Transmembrane Fragment of Band 3. Mild
1
0
,
1
2
,
1
4
,
__-.___ -..-._____.--
,
,
6
1
,
0
/
1c
Migration (cm) FIGURE 4: Fragmentation of band 3 by chymotrypsin digestion and separation of the major membrane polypeptides. Intact erythrocytes were treated with chymotrypsin and incubated with ~ - [ m e t h y l - ~ H ] methionine as described under Experimental Procedures. Control cells were incubated in parallel without protease. Acetic acid extracted membranes were electrophoresed, and the data were analyzed as described in the legend of Figure 3. In this particular experiment, a sample of 80 pg of total membrane protein (equivalent to 11.4 X lo7cells) was applied onto the control lane. The chymotrypsin lane contained a sample prepared from 1.4 times the amount of material (16.0 X lo7cells) as in the control lane. The 59 000-dalton fragment was found to contain 68% of the radioactivity in the intact band 3 polypeptide after the amount of material applied to each gel lane had been normalized according to the cell equivalence. It was calculated from seven different experiments that 60% 9% of the radioactivity in an intact band 3 polypeptide can be recovered in the 59 000-dalton fragment after chymotrypsin digestion.
*
chymotrypsin treatment of unsealed membranes releases the cytoplasmic domain into the soluble fraction, leaving two major integral fragments with molecular weights of 14 500-19 000 and 35 000-38 000 in the membranes (Figure 2; Steck et al., 1978; Reithmeier, 1979; Drickamer, 1978). With the exception of the N-terminal lysine residue, the sequence at the amino terminus of the 14 500-19 000-dalton fragment is identical with that of the 52 000-60 000-dalton tryptic fragment, and thus it appears that the Chz chymotryptic site is a single residue before the TI tryptic site (Mawby & Findlay, 1982). Chymotrypsin digestion of membranes that were prepared from 3H-methylated intact cells results in the disappearance of band 3 protein and the appearance of a radioactive 35 000-dalton membrane-bound fragment representing the C-terminal chymotryptic fragment. This digestion also generated a 15 000-dalton membrane fragment which respresents the 14 500-19 000-dalton fragment expected. This fragment had no detectable radioactivity (data not shown). We were surprised by this latter result, because this fragment would be expected to contain all of the methylation sites (estimated at 20% of the total methyl groups) common to the tryptic 52 000-60 000-dalton C-terminal segment and the chymotryptic 55 000-59 000-dalton N-terminal segment. Unless the presence of the methyl ester can allow protease digestion at sites not normally cleaved under these conditions (for example, at Ch3), these results suggest that the number of methylation sites in the central region of the polypeptide may have been
56
LOU AND CLARKE
BIOCHEMISTRY
Table I: Inhibition of Carboxyl Methylation and D-Aspartic Acid @-[3H]MethylEster Formation in Intact Erythrocytes Incubated with Adenosine and Homocysteine Thiolactone and Cycloleucine' pmol of D - A s ~ pmol of Me ester from band @-["]Me ester from 3b/mg of total membrane band 3/mg of total inhibitors added during incubation of cells protein membrane protein (A) none (control) 2.6 (lOO%)c 0.23 (lOO%)c (B) 5 mM each of adenosine and homocysteine thiolactone 0.14 (5.4)
